Keywords

1 Introduction

As carefully outlined thus far in the book, the space of global governance is constituted by various forms of knowledge practices and interdependencies between heterogeneous actors. Global governance is often described as a practice at the intersection of knowledge and policy (Haas, 2004). In this domain, these two spheres are particularly closely linked—where ‘power is a disposition (in the sense of ordering or controlling) that depends on knowledge’ (Adler & Bernstein, 2004, p. 294). The epistemic orders that dominate this space include technocratic modes of decision-making (Scheel & Ustek-Spilda, 2019) and quantification (Merry, 2016), which serve as ways of decontextualising knowledge spaces in order to create ‘the global’—both a community that can work together and the common issues on which this community can focus. Nevertheless, global governance is never fully devoid of the ‘local’—it is inherently a political space in which contextualised interests, actors and networks of power and influence overlap (Stone, 2019).

These spaces require ‘readily comparable and accessible knowledge’ (Rottenburg et al., 2015, p. 2). However, against the background of the highly fragmented structure of global governance a key question emerges: what forms and formats of knowledge fit these criteria? One approach is to focus on knowledge that has ‘universalizing’ qualities—such as numbers, standards and benchmarks (Timmermans & Epstein, 2010). These formats constitute global knowledge as they represent the ‘view from nowhere’ (Jasanoff, 2011)—knowledge that is context-less and as such is mobile and applicable across different boundaries. Nonetheless, focusing on the ‘view from nowhere’ presents only part of the picture—particularly in the context of the growing paradigm of participation and democratisation of global decision-making, as reflected in the SDGs (see Chap. 5). Instead, global knowledge is increasingly characterised by what Mike Hulme (2010, p. 559) referred to as ‘view from everywhere’: ‘knowledge which erases geographical and cultural difference and in which scale collapses to the global’.

Production of this type of knowledge is challenging as it spans multiple boundaries but also multiple knowledge orders. The ‘machinery of knowledge production’ (Knorr-Cetina, 1999) on the global level consequently encapsulates not the only production of research but—perhaps more importantly—unprecedented levels of coordination of actors, knowledge and practices. Within the SDGs, this machinery of knowledge production explicitly requires interdependencies between countries, International Organisations and supra-governmental groups—which we analyse as the second order of the epistemic infrastructure.

Against this backdrop, new forms of expertise emerge: as we have already seen, these include expertise in the production of narratives and visuals, the harmonisation of data and the promotion of participatory engagement and country buy-in. International Organisations are no longer just ‘knowledge institutions’ (Miller, 2007) producing numbers for global governance but are also boundary organisations (Guston, 2001), located at the intersection of different institutional, epistemic and political orders. As plentiful examples in this book have shown so far, the most important consequence of this evolution is the changing role of expertise: in this complex and fluid context, IOs are not only producers of numbers but rather coordinators of number production. They are ‘expert brokers’ whose role is not merely to produce knowledge but rather to create the conditions under which global knowledge can be produced.

Therefore, global governance as exemplified in the emergence of the Sustainable Development Goals is a unique space that is highly fragmented: hence, the processes of ‘unification’ (even if for short periods of time) of these diverse entities happen predominantly through knowledge practices. Consequently, the way of upholding the ‘global’ problems required in the SDGs is to create a common knowledge framework for knowing the problems. We have discussed this notion in Chap. 2; in this chapter, we turn to the actual role of actors who maintain these structures and practices. Numbers do not work on their own but rather they require a set of rhetorical and epistemic practices to make them operational. In this chapter, we turn to those practices undertaken by IOs to govern the SDGs and their key role in producing and maintaining the epistemic infrastructure.

The next section outlines the key debates in the literature on knowledge brokers and boundary work, which is followed by the analysis of the ways in which experts working in IOs become ‘expert brokers’. We unpack this concept by focusing on three types of boundaries that these actors navigate: institutional, epistemic and praxis boundaries. We conclude this chapter by outlining the conditions under which global knowledge for sustainability is constructed.

2 Boundaries and Bridges

In the analysis of expert brokers, one of the central points of analysis is the boundaries between different knowledge systems. The concept of ‘boundary work’ was introduced by Thomas Gieryn (1983) as an approach to identifying the difference between science and other areas of human activity. Science, as argued by Gieryn (1983), is not identified by any essential, inherent characteristic but rather is demarcated by the rhetorical work of different actors as means of securing influence and resources. On the most pragmatic level, such a division helps to share labour between science and policy and to assign responsibilities for different elements of the science-into-policy process (Huitema & Turnhout, 2009). On the more conceptual level, such divisions play a role in differentiating between the ‘technical’ and the ‘political’, and therefore acting as lines of demarcation between knowledge and politics, fact and value, objectivity and interests (Turnhout et al., 2008).

Other scholarship has focused not necessarily on the demarcation but also on the navigation of the boundary (Halffman & Hoppe, 2004), hence assuming the flexibility and hybridity of the boundary (Epstein, 2011). Navigation of the boundary is one of the key practices necessary for the uptake of evidence in policymaking—which has led to a growing focus on the actors ‘in-between’ knowledge production and policy. The scholarship has explored different ‘boundary bridges’ (Wenger, 1998 see also: Kislov, 2014): knowledge brokers, boundary objects and boundary interactions. Knowledge brokers—individuals (or organisations)—are entities whose goal is to link different communities, translating knowledge between them and building capacity for cross-boundary engagement (Bandola-Gill & Lyall, 2017). They work on the periphery of two different social settings and are charged with enabling the collaboration and interaction between them (McNie, 2007; Miller, 2001). In order to achieve this, they must be perceived as grounded in and legitimate in more than one area of practice, as such they are accountable to both sides of the boundary (Guston, 2001). Finally, boundary interactions refer to different forms of connections between people from different domains, for example—meetings, networking, shared projects and collaborations. These connections require diffusing the ‘models of knowledge’ (Lamont & Molnár, 2002) across different settings. Akin to Star and Griesemer (1989) view, the boundaries not only are markers of difference but also enable communication. Here, the focus is not on separating the practices but rather assuring their continuity across boundaries (Kislov, 2014). Building continuity across such boundaries is crucial to the functioning of the epistemic infrastructure as a whole.

Finally, these connections and continuities are often forged through indicators themselves, which serve as boundary objects and mediators in this global governance space. Star and Griesemer (1989, p. 393) describe boundary objects as those ‘which are both plastic enough to adapt to local needs and the constraints of the several parties employing them, yet robust enough to maintain a common identity across sites’. These objects have different definitions in different social worlds, yet they are recognisable across these disciplinary or governmental boundaries.

3 Expert Brokers: IOs and New Forms of Expertise

Perhaps the most significant finding coming out of the METRO project was the fact that, despite an undeniable commitment to governance by numbers, IOs did not see themselves as ‘producers’ of statistics. Rather, they saw themselves as ‘links’ between a variety of actors, navigating the complexities of knowledge production on the one hand and the politics of number-making on the other (see Bandola-Gill, 2021). This does not mean that the experts completely abandoned their statistical and economic training—quite the opposite, they perceived their work to be knowledge-intensive and requiring high-level skills. However, the central practice for experts working in IOs was not the production of knowledge but the coordination of knowledge production.

What we observe in this setting is an emergence of a specific form of expertise—one that surpasses producing and disseminating knowledge. Instead—this work requires creating specific knowledge environments in which knowledge is agreed upon and produced. This activity is not just translation (as discussed in the literature on knowledge brokers outlined in the preceding section): what these actors are doing is much more complex and involves creating the conditions for knowledge production through setting standards, coordinating their application and—finally—navigating the communication between these complex networks of actors. This process of coordination is complex, requiring bridging multiple boundaries at once—including boundaries between different institutional settings, epistemic boundaries between different forms of knowledge and differences in practices. These three types of boundary practices will be discussed in the remaining part of the chapter.

3.1 Brokering Knowledge

The first dimension of the process of creating global knowledge involved brokering knowledge. Even though experts working in IOs saw their role as predominantly supporting evidence-based policymaking—linking quantitative evidence with policy and practice—this process was more complex than translation between knowledge and action. Instead, the process involved navigating between different knowledge systems—from scientific, political and economic to local and indigenous and practice-based. This translation was multifaceted and required skills and integration of all different epistemic orders at once, which was challenging:

This is a very country-specific question, it’s just very specific to where are the levers, where’s the interest in government, who are the partners? You need to be able to work within that realm to find your way to what is going to be most effective. So I think we just want to find a balance [with strict guidelines]. But it’s really encouraging that balance, but within some parameters which fall within: ‘yes this is child poverty measurement’. (UNICEF, 3)

On the country level, this translation between knowledge of different epistemic standing (c.f. Bandola-Gill, 2019), for example, between experiences of poor people, statistical and economic methodologies and political focus on ‘doability’, required embedding the numbers in other elements of the existing knowledge system—for example, translating them into storylines (see Chap. 4), linking them to existing agendas, policy solutions, etc. The key epistemic challenges the brokers must navigate are the different understandings of ‘numbers’ between experts and policymakers, including the methodological and technical standards and political context:

A lot of times I feel, and this is something that we’ve talked about with our teams very frequently, that our role is not just to provide the best possible technical advice but it’s also to work with our counterparts to think about how can they communicate that in simple understandable terms to the population or to other parts of the government who may not be experts in poverty measurement. So yeah, just being aware of what’s the political context in which these changes are taking place is very critical. (World Bank, 12)

Here, opening the channels of communication was the key practice. As recalled by one of the UNICEF members—‘the worst thing you can do is to just drop the report on their lap’ (UNICEF, 7). The role of the brokers was to get the policymakers involved in setting the priorities, choosing the dimensions and shaping the measurement process. Importantly, another key area of expert brokers’ work was to keep the policymakers abreast of the numbers that are coming—particularly if the numbers were negative. As such, the co-production of quantified projects—in which the government, IOs, statistical offices and civil society representatives worked together to produce the measures and the reports together—became a space of boundary interaction in which these communities, previously fragmented, were becoming a community of practice.

On the global level, the process of consensus-making and coordination between different ways of knowing was complex. One specific site requiring navigation between different ways of knowing in order to achieve the ‘view from everywhere’ was a negotiation around the SDG reporting (Jasanoff, 2017). In the case of global poverty, this process was particularly complex, as the interviewees saw poverty as a situated and country-defined issue. This was an area in which different knowledge systems were competing. On the one hand, different countries as well as IOs were proposing poverty measurement more closely aligned with their understanding of poverty. Unsurprisingly, different organisations called for the inclusion of dimensions of poverty aligned with their organisational remit. These additional dimensions were not inconsequential as they opened up multiple possibilities for the custodian agency in charge of the reporting for this indicator. For example, UNICEF’s written comments in the consultation on the SDG indicators submitted for the first meeting of IAEG-SDGs as Indicator proposals received from agencies (IAEG-SDGs, 2015) suggested the addition of child poverty but also suggested themselves alongside the UNDP as custodians of the target (based on the scale of coverage by both organisations):

Proportion of children living in multidimensional poverty. This indicator is expressed as a percentage. Deprivation dimensions and indicators should be based on internationally agreed standards and definitions. Deprivation dimensions include inter alia: nutrition, education, health, housing, water and sanitation. (UNICEF)

The International Labour Organisation (ILO) requested employment status and saw themselves (alongside the WB) as the custodians. The International Fund for Agricultural Development (IFAD) called for an addition of disaggregation between rural and urban areas. The working group on disabilities wanted a disaggregation between able and disabled. The International Finance Corporation (IFC) asked for an addition of another indicator to this target:

Percentage of population using banking services. Please disaggregate by gender. (IFC)

Correspondingly, the UN WOMEN requested women-centred indicator:

Proportion of people who have an independent source of income by sex, age and source of income. (UN WOMEN)

These debates went beyond the content of indicators but also accounted for methodological standards. The best example here was the position of EUROSTAT:

MPI should have the form of the EU 2020 poverty and exclusion indicator: if a person suffers from any dimension of poverty s/he has to be considered ‘poor’. Statistical compensation of poverty dimensions against each other should not be allowed. Dimensions should be oriented towards basic needs: enough clean water, enough healthy food, clean air, shelter, security, basic education. (EUROSTAT)

On top of these debates over the understanding of poverty through disciplinary and organisational lenses, different countries were opposing measures that were not aligned with country-level definitions, as one interviewee put it:

For multidimensional poverty, I think the problem is it’s more acute, because countries are very different, so depending on how you choose the dimensions, some dimensions would be relevant in some context and not relevant in other contexts, so when you take dimensions or indicators that, for example, look at access to technology or access to a TV, which was the case in some of these indicators, you will find that in some societies everybody has that basically, it’s a given, whereas in some societies there’s still a significant gap in access to that kind of technology. So, if you use that in a setting where everybody basically has, let’s say everybody has internet and broadband, you can’t use that as a measure of poverty, because then nobody would be poor in that country. […] That’s one of the reasons why you had this pushback by some Member States who didn’t want to have a global definition that would determine the number of poor people in their country and that may not necessarily reflect the experience of poverty in their countries. (UNDP, 4)

It soon became clear that achieving the perfect consensus (‘view from nowhere’—Jasanoff, 2017) was not achievable in this context, At the second meeting of IAEG-SDGs, the representative of Mexico’s CONEVAL (The National Council for the Evaluation of Social Development Policy) advocated for a country-led indicator:

Based on the foregoing, it is proposed that the indicator for Target 1.2 traced for Objective 1 could be stated as: ‘By 2030, reduce at least by half the proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions.’ This proposal comprehends two essential aspects of Target 1.2: it i) recognizes the multidimensionality of poverty, and ii) favours each country with their specific needs. Also, allowing each country’s particular definitions would take into account the specific contexts of poverty and the availability of sources of information in each one of them. (Mexico’s country statement)

This proposal was challenged by Sabine Alkire—who works for OPHI and is one of the co-creators of the MPI—in her ‘Academic Stakeholder Statement’ in which she argued that only a global MPI would provide a comparable measure. She claimed that relying on country measures will not benefit the poor—because the national measures are not comparable and consequently in some countries, this approach will result in a very limited anti-poverty action.

In the end, as we have already seen in Chap. 5, the approach discussed by Mexico was indeed implemented and eventually, the country-led measurement was introduced to the framework. Therefore, in the end, the consensus in this case led to prioritising country-level data—or ‘view from here’ rather than the preferred option of global-level data. However, the alignment of actors was seen as a priority over finding ‘perfect’ data solutions.

3.2 Brokering Practice

The second boundary that was ‘brokered’ by the experts working in IOs was one of practices—including workshops, training sessions, online learning and handbooks aimed at capacity building. These strategies were not just aimed at creating specific knowledge repertoires for the stakeholders to use but they have a more important role in creating—what Galison (1997) called the ‘rules of exchange’ within this particular ‘trading zone’ (Collins et al., 2007). These rules of exchange allow for coordinating and collaborating between different actors without necessarily requiring the consensus.

For example, in the case of global poverty, this meant that economic and statistical methods were constructed as a standardised framework for exchange—of ideas, priorities and items on the agenda. This was not a mere capacity building exercise, as it was providing language through which the participants could communicate—the participants were expressing their ideas and priorities in the language provided by the methodological framework. As we discuss in Chap. 4, this common language provides a narrative for the global public policy of sustainability, providing a ‘glue’ for the epistemic infrastructure. Interestingly—and relatedly—the process of developing capabilities during the workshops was not completely open—the boundaries were established by the methodology itself, even though there was a level of flexibility:

Stakeholders, I mean, you have to facilitate to them. Of course, because this [multidimensional poverty] is a new concept to many of them. In a way, we try to be very honest, we try to make the stakeholders to believe in us and try to make them reach a consensus that we want them to reach. Just to put it very, very frankly. Of course, we design all these workshops and materials, we give presentations, we convince them. This is firstly important. Secondly, the methodologies we are using are very relevant and they’re technically sound methodologies. And then we try to engage them to identify what are the most important dimensions, but even for that, we need to come up with a list. As far as I recall, we did not let people brainstorm. We already had a list. Then people went through the list and say, okay, they select some. Although, I think in the case of Botswana, actually, they added something. […] As you can imagine for African countries, crime safety is a problem, so they want to add a crime safety dimension to it. And we have had no problem with that. If they think this is an important aspect of possible deprivation for children. So, we added that. (UNICEF, 10)

However, IOs’ experts were aware that capacity building—with its designated roles for certain actors—was not necessarily aiding progress towards achieving the SDGs:

Policy’s really hard, because even if we think of a country with just a well-functioning data architecture and system in place, if I think of the US, it has excellent data sources, government technicians are very strong, collaboration between academics and government and policymakers are probably some of the best in the world. And yet, I’m not sure I could point to specific examples of data that’s been collected that has directly influenced policy. (World Bank, 2)

In many cases, capacity building was not only a matter of knowledge creation but also a matter of getting all the actors to get the common ground and thus enabling communication. As we discussed in Chap. 3, statistical capacity development made explicit how central creating these conditions of knowledge production is to the functioning of the epistemic infrastructure of the SDGs. By mapping out and facilitating the materialities needed to expand, create and maintain data and statistical systems, IOs function as brokers for the entire system.

4 Brokering Global Public Policy: Boundary Work as Infrastructuring

Finally, we turn to the brokering role of numbers themselves. The predominant outlook in the literature on indicators (Merry, 2016) posits numbers as effective forms of both communication and governance as they create mobile knowledge which travels across different contexts. This aspect of numbers is indeed important and evidently, the SDG reporting framework draws heavily on this ‘universality’ of numbers, allowing for articulation of common goals within one monitoring scheme. Nevertheless, our findings point to another, more situated role of quantification—namely the ability to create fora for debate and consultation amongst multiple stakeholders. This ‘new’ role of numbers in this context was made possible by the acceptance of quantification as a governing paradigm (Tichenor et al., 2022), as best illustrated by the fact that, despite ranging views of SDGs themselves, none of our interviewees questioned the idea of measurement itself. One of the interviewees summarised this paradigm (attributed to Andreas Schleicher, Director of Education and Skills at the OECD): ‘measuring the pig doesn’t make it fatter, but at least it can tell us if it’s overweight or underweight’.

At the same time, the interviewees were unified in their perception that it was the process of producing the numbers—negotiation, consultation and consensus-building—which was as important as the numbers themselves. This unprecedented focus on the process of collaboratively producing numbers in the global governing spaces highlighted two interlinked qualities of numbers in this setting. First, numbers do not exist independently of the social actors producing them. In the context of the SDGs, this increasing coupling of actors and numbers led to the rising importance of the process of selection and coordination of multiple stakeholders. This makes the role of a broker not only one of a connector (Meyer, 2010) but also one of an architect—by shaping who gets invited to the process and who has a say (as will be illustrated in the example of meetings)—key to constructing the larger epistemic infrastructure of the SDGs. Second, numbers do not only represent the issues and communities but they also create them. As such, the process of quantification emerges as a process of ‘stakeholdersation’ (Metzger, 2013)—where indicators transform groups of actors into stakeholders of issue that are being measured. These two qualities of ‘brokering connections’ will be discussed in the remaining part of this section.

4.1 Numbers as Mediators

One of the key effects of the indicators, as identified by the vast majority of interviewees, was that they enable collective deliberation and decision-making. This ability to ‘bring people into the room’ (UNICEF 2, 4) was seen as central in this process as it opened up—often cross-sectoral—channels of communication and constituted a strategy for establishing common ground. This quality is central in creating cross-boundary initiatives (Bechky, 2003) and in the context of the SDGs it was achieved through quantification. This was particularly important for settings where the group structure was difficult to be maintained. One such area is poverty which is a multidimensional phenomenon (Atkinson, 2019). This makes it a difficult policy concept, as the policymaking happens within organisational silos.

The development of multidimensional poverty measurement enabled the cross-sectoral engagement on a scale that surprised even the creators of the measures. As recalled by one of the interviewees, the most unexpected consequence of the development of the multidimensional poverty measurement was the fact that it brought various groups of stakeholders together and enabled deliberation. The interviewee argued that the indicator brought together departments which usually do not work together, even at the national level:

Often ministers compete with each other: ‘I want to be the most important minister, I want the most budget’. And when you have an MPI and the minister sit at the table then you can’t move the MPI down the field single-handedly, you need a team. And so, you need to kick from the minister of health, kick from the education and together, as a team effort, they can move the ball on property. And they learned that. So actually, you might say ‘I’m sending an education indicator to the education minister responsible for that’. But then she will say, ‘I cannot make my education goal without the other ones’. So actually, they learn about what the others are talking about, how we need to integrate the policies, but they learn it from each other. And we’ve seen it in the number of governments where ministers actually reach a common understanding. Well maybe there they compete in other spheres but when it comes to poverty the moral imperative is so great that then they said we going to cooperate. (OPHI, 1)

Therefore, the indicators themselves have specific ‘constituencies’ (Voß & Simons, 2014)—sets of actors which emerge to maintain or develop specific tools as modes of governing. Even though this quality of the multidimensional poverty indicator was initially seen as an unintended consequence of the methodological innovation, it soon became common practice to mobilise it strategically. The central affordance in which indicators as material mediators shaped the practices was through their impact on practices. The practice of producing the indicator not only focused on the indicator itself but also created specific knowledge/knowing spaces:

I think my observation here is we need to engage people at the right level, as high as possible, and do good facilitation. Don’t just let them brainstorm. It’s not going to be helpful because they don’t even have the basics of what we are going to do. But do give them enough space to talk about what they think would be most important. […] Also actually, we also involved—I think we involved the academia from the country. So, we involved some university people to be with us in the facilitation, so that we are not really coming up with crazy things that are not relevant. But we make sure that these are relevant, and they’re endorsed by local researchers, as well. I think that would be important. (UNICEF, 7)

Even though the process of developing the indicators supported the creation of constituencies, the process of generating global knowledge required more extensive coordination of different actors. The process of ‘production of numbers’ for global governance is in fact a complex navigation between these multiple actors who are producing the measures:

I think it boils down to how you can work with your counterparts to, in a way, get them to accept that this is what the evidence says, but also understand that they don’t only have technical considerations, they have other considerations and work with them in terms of well how can this be useful to you. Maybe it’s not the news you expected, but it’s still the news, so what does this mean, you know, is there something that can be done proactively about this and so on. But it’s not always easy and we have faced situations where the government didn’t want to publish the numbers and the numbers have been not published or have been published with a delay. (World Bank, 9)

On the global level, as we discussed in Chap. 2, these indicators mediate the frameworks and needs of other diverse groups of actors. As IOs create spaces for deliberating the international concept and the measurement tools for each indicator—as part of the required protocol for establishing or reclassifying an indicator for the framework set by the IAEG-SDGs and the UN Statistical Commission—civil society groups, member states, regional organisations and IOs provide both conceptual and methodological feedback to shape this unifying object of concern for the specified policy arena. As was the case with the indicator on governmental migration policies outlined in that chapter, almost 300 governments, International Organisations, civil society groups and academics were brought together by the process, allowing the indicator to become the common language spoken within such a vastly different group of stakeholders in shaping what is knowable about migration. Each indicator does this work of brokering diverse actors and perspectives in the infrastructuring of global public policy.

5 Conclusion

This chapter began with the acknowledgement of the complexity of global governance. It has become increasingly fragmented, and this fragmentation necessarily requires a form of linking, of collaboration and of creating common deliberative spaces. Consequently, governing these spaces requires creating specific conditions for knowledge production which become constructed not only by a specific set of indicators but also by the act of governing itself—the ways in which experts become not only producers of knowledge but also—or even exclusively—knowledge governors. They do not produce numbers—they broker connections, build consensus and work towards the shared meanings, rules of exchange and material artefacts.

The scholarship on knowledge brokers sees them predominantly as actors ‘in-between’ whose role is supporting evidence translation between research and policy (Meyer, 2010; Turnhout et al., 2008). Even though this was one of the types of practices that the experts in IOs engaged in, their role was in fact more extensive. By ‘bridging’ different groups of actors, knowledge and practices, the expert brokers not only engaged in linking different epistemic and institutional orders but rather supported creation of unique forms of ‘global knowledge’. The process of creating the conditions for knowledge production was not merely an act of translation but rather an act of creating new epistemic environments which then determined the forms of knowledge that were possible to produce. As such, these brokers acted as institutional filters, transforming a multiplicity of ways of knowing, organisational structures and political priorities into common epistemic and political frames. This was particularly important in a context of the high fragmentation of global governance—the role of brokering work was to create conditions for unifying of knowledge, without necessarily universalising it.

The key strategy here was to create specific ‘fora’ for engagement—for example, through meetings, indicator development or training workshops. The role of brokers was to bring different actors together to deliberate and build consensus. It was in these fora where ‘the scale collapses to global’, even if temporarily. The production of global knowledge was therefore a process of bringing actors together and letting them go back to their ‘main’ institutional settings and through this process allowing for the co-existence of multiple, sometimes contradictory, knowledge systems.